An Open World (Cup)

Written by glenn mcdonald Tuesday, 11 May 2010 15:12

There are lots of sources of good historical World Cup information, including FIFA's own site, Wikipedia and worldcup-history.com. Any particular known World Cup fact you want, you can probably find on any of these sites.

But on all of these sites, as with most websites in the world, when their facts run out, that's the end. That is, their set of answers is fixed. If your question matches an answer they thought of ahead of time, then you can have it. But you can't ask any new questions. You can't explore variations on their current questions, or see holistic views of anything they itemize (or vice versa), or even debug their answers if you wonder about them. You can't turn their data around to any perspective they didn't provide.

There's a simple reason for this, of course: that's what most tools for data publishing are built to do. They take databases and produce sets of static reports. The better tools link those reports, at least a little bit, and some of them even help you build something with at least a few moving parts: tables the reader can re-sort, maybe.

In Needle, all the parts move. Needle's version of World Cup History is not a sheaf of reports, it is an explorable dataspace. The pages you move through in Needle are the paths and properties of the data itself. You can see any perspective that exists in the data, any subset, any sort, any tabulation. I spent a few hours answering questions I had, and you can click down the index on the left to see what I found, but if I didn't answer your question, Needle will let you try to answer it yourself. If there aren't enough widgets for what you want to know, there's even a query language to let you ask more or less anything the data can express, and if you still aren't happy, any data you see can be exported for processing elsewhere.

This is our idea, or at least our progress towards our idea, of what openness should mean in data-publishing.

< Prev		Next >